The weighted incremental norm approach: From linear to nonlinear \(H_\infty\) control (Q5947629)

The paper shows that a possible way to extend the \(H_\infty\) approach to the nonlinear context can be the weighted incremental norms. The main conclusion is that the incremental framework allows one to take into account not only the classical linear requirements such as robust stability with respect to unstructured uncertainties, attenuation with respect to output perturbations and sensitivity, but also specific problems associated with the nonlinear nature of the plant such as initial condition uncertainties and steady state behaviors with respect to specific classes of inputs (constant or periodic). Weighted induced norms, namely \({\mathcal L}_2\) or incremental gains, are used to handle, in a weak sense, constraints on the behavior of the closed-loop system. Some limitations of the \({\mathcal L}_2\) gain approach are pointed out, and then it is shown how, under suitable assumptions, incremental boundedness allows one to bypass these limitations.